Detection and Classification of Cognitive Distortions: A Literature Review

Various machine learning and deep learning approaches have recently been applied to detecting and classifying cognitive distortions (CD). However, there are several challenges in detecting and classifying cognitive distortions. This paper outlines current research in CD detection and classification, challenges and problems. The first challenge lies in the limited availability and accessibility of public datasets. Another issue relates to CD domains, which involve short text data that may only partially represent a particular CD category. The performance of CD detection model has a relatively good value, while in general, the classification process has not shown promising results. One model that works well and consistently is BERT, both in terms of vector representation and as a classifier. In addition to the model, dataset availability and reliability remains an important issues to address.